智慧型多媒體資訊處理系統的研究(I)

全文

(1)

(2)

(3) The Study of Intelligent Multimedia Information Processing System(I) .

(4)

(5) . !" #$%&'()*+ .

(6)

(7)

(8)

(9) !"#$%&'() "*+,?ghijk>l 4mnompq,r?6G stu2 v$&wxyz,{| g}~QD +,6?69V ab/6% l4ab,l4ab #no,

(10) Q WO<01 ¡ K¢c00£¤¤¤¥M,m /7¦§m¨? WWW N©ª2 $%&«)"*+,-. ;<0 1DEFab#¬/7gh i?© query "{D©®Z;< 01¯°±ZO<01,GO<01 ®Z/²³´µ¶¯·°/¸´ µ¶, ¹º»?O<01A¼½¾ %¿À ghi?© ÁÂÃÄÅ ",E6"WXY UZ% & [\] DEF^_` a b 3 4 K Intelligent Multimedia Information Processing Systems, IMIPSM, WÆÇÈÉQZ J D2 9:^_àb34[5], ÊË|34ÌÍ¿À

(11) ÎÏÐbÑÒ ,.

(12)

(13)

(14) !"#$%& '()"*+,-.#/012 3 45)6/7#89:2 ;<01=>?@AB5 C&01 DEF,G?6HI J KLLLMN O<012 -.P/QRSTU V,WXY UZ%& [\] DE F^_àb34 K Intelligent Multimedia Information Processing Systems, IMIPSM, c^_àb34d[ \] dO<012 ef Abstract Retrieving multimedia information efficiently among the world wide web becomes more and more demand. Many images search engines have been developed, so far their retrieved image were based on the image database constructed by themselves, and the developed of video system are seldom to be announced. In this project, We intend to design an intelligent multimedia information processing system (IMIPS) based on neural networks.. Keywords: IMIPS, neural networks, video search engine 1.

(15) IMIPS ÌÍ¿ÀÓÔ,ÇÈ IMIPS ÕQÕQ Ö"×&ØÕÕ ÙD Úzi (Server) ÛØÕ 1. Video Spider 2. Keyframes Selector 3. Feature Extractor 4. DataBase Management System ÜÝ(Client) ÛØÕ 5. User Interface 6. Query Processor . 2. maintain-spider = travel-spider ëì N íî (URL) ïmaintain-spider ðëì ñò#ó íôõ& íö÷ /äOø")ù?ú íûüÕý þ9:9ø¸, Keyframes Selector Ö"ÎÏ$ Spider 2£©ª.ab,6© 9>;<01"{|?6 ©ÇÈ¹ºª%ab Ï , Ç È u % ¼ Y (keyframe)j © P

(16) t³´

(17) ªAB u,ÇÈ ©] MPEG ju ¼Y ,6©] MPEG W 9«§= I-frame hª ¼Y,®|@¯ÇÈ° > I-frame ´jU ¼ Yû%g9 ,6 Vt6 WWW N©D MPEG ß{ |ÇÈ ¼Y u'( h6&H ,§!©D MPEG ß Ç È " h # / # =ªQ MPEG ß[2],. Feature Extractor DÌQ IMIPS )" Øê@%,=ÎÏ$©$t% ©AB uUj! Feature Extractor ¬ «©AB uUjðC &34&¦%å',ÇÈ$ (C);<01*+Q,-./ VË|×ØÕÎÏÐbÞÒ, -Ì0U12AB,ÇÈ. 34ABh Multiple-Instance 56

(18) Intelligent Video Spider Ö"ÎÏ h EM 78U9W AB[3], $ WWW N2£©ßà,DZá Database Ö"$Õý:9Ï/ âã WWW N©9äOSpider ¹ 7;b,&34/<u9ó ºå¸ WWW 2 ©9,/7 Õý9=>?G ¦ WWW N2 $ØÕãæç ¹º@jÏDAÎÏB4 è,W34 Spider Ö"ÕQé&Øê ôC/:j;bj [1] tu¸QD34ëDDZ E 1. travel-spider 2.

(19) , WXY)/7uh©9 Õýab©9

(20) ¨2 H [,;$%[`2 ef 9£$¬r- User Interface Ö"$ IMIPS Pgh 2 ef£9y,-. iQRST,=UX)V6W 2 efØê$ DÖ© ghi«÷XP IMIPS QR Y 2 ef·^Pfgg/«P Z[$ User Interface UX-\,E 12 efGÇÈ -\$ghighøÇÈ34 6&-\ÇÈ]?I& query =hó!"ôDZ&-\ AB½¾^9W9:[3]ghi ÇÈÕý0wö÷hóabôË 9W9:V09Wª query, :1/ ¡¢ø0wF?0wÏZ£ ¤,zj©9¥ Query Processor ð$ User Interface R6 N¸}øW34¦Wg _ûj query ]ø%ab u hihó{!", query A B , ` _ û Ú z i Û Similarity Comparator ª>[3], . 34ëDÇÈh9:A¼? "

(21) tÏ/7;bFBG "

(22) hHI&[Ùgh%&J ¾ table jKL¸M NOU9[1],. . |¯DZ¡a IMIPS ab \2^!! "The study of intelligent spider for video search engine",

(23) bÇÈ Java Applet ½À%&[ , June 1999 c©dH efghijkH ljkimXnodpqdr \3^!! ,"The Study of Scene Change Detection ¥sÏt,ghi|H ?6 and Moving Object Detection in MPEG Video",

(24) , June 2 øuhóvw¡Nx

(25) yµO 1999 elzj={ghihï|$}ø| 34 IMIPS 2 hÏû% \4^!! ,"Image classification and indexing by jkh[4], Multiple-Instance Learning",

(26) , June 1999 \5^!! ,"A SMIL-based Authoring Environment for Internet Video",

(27) WÆD|XY~%Æ,WÆ , June 1999 ÇÈáÏZ%& O<©D2 j 2 ef`[5],V \6^!! The Intelligent Multimedia Information Processing System (IMIPS), “http:// www. csie. ©9F199 nctu. edu. tw/~ nnlab/ Demo/ IMIPS”

(28) G§-. N©9 =«rûù j N©9 9%9`G2 ef#©%9

(29) ¸"2 efjghi ó

(30) uh!" 3.

(31)